Ixx PEOCEEDI>'GS OF THE GEOLOGICAL SOCIETY. [vol. Ixxiii, 



lime and magnesia. As regards the mechanism of such separation, 

 Bowen lavs stress especially upon the sinking of crystals in a mass 

 which is still mainly liquid. Simple physical considemtions, how- 

 eyer, tell us that a great liquid or mainly liquid reservoir within 

 the earth's crust can haye only a temporary status. It is doubtless 

 incidental to the existence of much more extensiye tracts of the 

 lower crust in a semi-liquid state, that is, as a liquid crowded with 

 crystals or a crystalline fabric with interstitial liquid. Such a 

 condition must be postulated beneath any region which is the 

 theatre of long-continued igneous activit\% and a mass so constituted 

 possesses remarkable properties. In particular, it is very sensitiye 

 to unequally distributed stress, the liquid part tending to be driyen 

 from places of greater to places of less stress. When in such a 

 region an unequal disti'ibution of crustal stress is maintained suffi- 

 ciently long, there will be a continued flux of the interstitial 

 magma, which will be driven out from areas of special disturbance 

 to accumulate beneath areas of relatiye quiescence. If, in addition 

 to this lateral displacement, the magma works its way into the 

 upper crust or to the surface, highly alkaline rock-t^^pes may result. 

 Otherwise, and more generally, the magma so displaced will go to 

 impart a more or less marked richness in alkali to the crust beneath 

 the undisturbed areas, and the effect may be seen only at some later 

 epoch and in a less extreme manifestation. 



The doctrine of two classes of igneous rocks, alkaline and calcic, 

 liaving a significant geographical distribution in relation to the 

 great tectonic features of the globe, has, then, a certain justification. 

 It does not, of com-se, imply any shar23 diyision ; and perhajDs the 

 more philosophical conception is that of two opposite petro graphical 

 poles, towards which igneous rocks tend as a result of primary 

 differentiation. There is naturally some complication introduced 

 by subsequent processes, giving rise to the great diversit}^ of igneous 

 rocks known to petrograj)hers ; and these later processes may some- 

 times obscure, as regards individual rock-t^pes, the primary 

 characteristics. 



To illustrate these general remarks, and to put to the test the 

 element of hypothesis which they contain, I propose to recall 

 rapidly the leading events in the igneous history of the British 

 area. We are met at the outset by a difiiculty which arises from 

 lack of full knowledge. There is only one part of the country, 

 namely : the Scottish Highlands, where any large expanse of the 



